SOME ASYMPTOTIC PROPERTIES OF VARYING KERNEL DENSITY ESTIMATOR

Robert Mnatsakanov
Department of Statistics
West Virginia University

Abstract

        In this talk we introduce a new nonparametric estimator for probability density function defined on the non-negative real line. Our construction is based on the inverse gamma density function used as a kernel. It is shown that proposed estimator achieves the optimal Integrated Mean Squared Error (IMSE) within the class of non-negative estimators and is free from edge effect. The L1-consistency of the proposed estimator is established as well.